RuozhiBench: Evaluating LLMs with Logical Fallacies and Misleading Premises
This addresses a critical gap in assessing LLM reasoning robustness for applications requiring reliable logical analysis, though it is incremental as it focuses on a specific evaluation dataset.
The paper tackles the problem of evaluating large language models' (LLMs) ability to identify logical fallacies and misleading premises, finding that even top-performing models like Claude-3-haiku achieve only 62% accuracy compared to human performance of over 90%.
Recent advances in large language models (LLMs) have shown that they can answer questions requiring complex reasoning. However, their ability to identify and respond to text containing logical fallacies or deliberately misleading premises remains less studied. To address this gap, we introduce RuozhiBench, a bilingual dataset comprising 677 carefully curated questions that contain various forms of deceptive reasoning, meticulously crafted through extensive human effort and expert review. In a comprehensive evaluation of 17 LLMs from 5 Series over RuozhiBench using both open-ended and two-choice formats, we conduct extensive analyses on evaluation protocols and result patterns. Despite their high scores on conventional benchmarks, these models showed limited ability to detect and reason correctly about logical fallacies, with even the best-performing model, Claude-3-haiku, achieving only 62% accuracy compared to the human of more than 90%.